Environ. Sci. Technol. 2006, 40, 5937-5942
Influence of pH on Plutonium Desorption/Solubilization from Sediment D A N I E L I . K A P L A N , * ,‡ BRIAN A. POWELL,§ LEO GUMAPAS,† JOHN T. COATES,† ROBERT A. FJELD,† AND DAVID P. DIPRETE‡ Savannah River National Laboratory, Aiken, South Carolina, 29808, Lawrence Livermore National Laboratory, Livermore, California, and Department of Environmental Engineering & Science, Clemson University, Clemson, South Carolina, 29634
Aqueous Pu concentrations and oxidation state transformations as a function of pH were quantified and compared between sorption/desorption studies and literature solubility values. When Pu(V) was added to a red subsurface sandy-clay-loam sediment collected near Aiken, South Carolina, 99% of the Pu sorbed to the sediment within 48 h. Throughout the study, g 94% of the Puaq remained as Pu(V), whereas e 6% was Pu(VI) and e 1% was Pu(IV). This is in stark contrast to the sorbed Pu which was almost exclusively in the +4 oxidation state. The fraction of aqueous Pu (Puaq/Pusolid) decreased by >2 orders-ofmagnitude when the contact time was increased from 1to 33-days, presumably the result of Pu(V) reduction to Pu(IV). The desorption studies were conducted with a sediment that had been in contact with Pu (originally as PuIV(NO3)4) for 24 years. At near neutral pH, a decrease of 1-pH unit resulted in almost an order-of-magnitude increase in the concentration of Puaq (7.5 × 10-10 M at pH 7 and 3.6 × 10-9 M at pH 6). Similar to the sorption experiment, g 96% of the Puaq was Pu(V/VI). The Puaq concentrations from the desorption experiment were similar to those of the Pu(V) amended sorption studies that were permitted to equilibrate for 33 days, suggesting that the latter had reached steady state. The Puaq concentrations as a function of pH followed near identical trends with literature solubility values for PuO2(am), except that the desorption values were lower by a fixed amount, suggesting either Pu sorption was occurring in this sediment system or that a more crystalline, less soluble form of Pu existed in the sediment than in the literature water-PuO2(am) system. Based on Pu sorption experiments and measured sediment surface charge properties as a function of pH, the latter explanation appears more likely. pH had a more pronounced effect on solubility and Puaq concentrations than on sediment charge density (or Puaq oxidation state distribution). Slight changes in system pH can have a large impact on Pu
* Corresponding author phone: (803) 725-2363; fax: (803) 7254704; e-mail:
[email protected]. † Clemson University. ‡ Savannah River National Laboratory. § Lawrence Livermore National Laboratory. 10.1021/es060523s CCC: $33.50 Published on Web 08/29/2006
2006 American Chemical Society
solubility and the tendency of Pu to sorb to sediment, thereby influencing Pu subsurface mobility.
Introduction Plutonium can exist as Pu(III), Pu(IV), Pu(V), and Pu(VI) in the same natural groundwater sample (1). This diversity of oxidation states greatly adds to the complexity of the chemical speciation of Pu in both the aqueous and solid phases and, therefore, impacts Pu chemical behavior. Pu(IV) forms strong hydroxide complexes and precipitates as Pu(OH)4(s) in neutral pH solutions (2, 3). The aqueous-phase concentration of Pu(IV) is believed to be controlled by formation of Pu(hydr)oxide solid phases. Typically, solutions with high Eh and pH values tend to form Pu(V) and Pu(VI) (1). Pu(V) has been shown as the stable aqueous-phase oxidation state in dilute salt solutions (1). Keeney-Kennicutt and Morse (4) speculated that Pu associated with sediments is typically found as Pu(IV) and that soluble Pu is present as Pu(V) or Pu(VI). Plutonium oxidation state is directly influenced by the Eh and pH of a system (or system acidity). In aqueous systems, these parameters are closely related to one another; a change in one leads to a change in the other. For example, in the reduction of PuVO2+:
PuVO2+ + e- + 4H+ T Pu4+ + 2H2O
(1)
Four protons are consumed each time an electron is consumed. Consequently, any changes in the redox status produces changes in the proton concentration. The extent that the measured pH changes, depends on the system buffering capacity. Conversely, eq 1 also illustrates that changes in proton concentration can promote oxidation state changes in Pu, providing electrons are available. If all other conditions remain the same, an increase in proton activity will promote reduction in eq 1. System acidity also has a profound influence on the solubility of Pu species (5). This is demonstrated by the greater than one stoichiometric value for the hydroxide values in the following Pu solid-phase solubility equations. System
Pu(OH)4(am) T Pu4+ + 4OH-
(2)
PuO2-2H2O T Pu4+ + 4OH-(3)
(3)
PuO2(OH)2(s) T PuO22+ + 2OH-
(4)
pH also greatly affects complexation and hydrolysis (5, 6), both of these reactions in turn may influence geochemical behavior (4, 7-9). Several oxide minerals have been found to mediate either reduction or oxidation of Pu (4, 7, 10-14). Pu(VI) kinetic experiments with manganite (MnOOH) and hausmannite (Mn3O4) indicated that the surface complexation of Pu occurs over the first 24 h of contact (12). The sorption increased with pH beginning at pH 3 until it reached a maximum value of 100% at pH 8 and then decreased over the pH range from 8 to 10. Surface-mediated reduction of Pu(V) was observed for hematite (R-Fe2O3) and goethite (R-FeOOH) at pH values >4.5 (11). At pH < 3, no sorption, and therefore no reduction, of Pu(V) was observed in these minerals. For hematite, Pu(V) adsorption was the rate-limiting step in the adsorption/ reduction process. Morgenstern and Choppin (15) showed oxidation of Pu(IV) on a synthetic manganese dioxide (MnO2) VOL. 40, NO. 19, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5937
using solvent extraction. Duff et al. (16) used X-ray Adsorption Near Edge Spectroscopy (XANES) to show oxidation of Pu(V) on a naturally occurring manganese oxide mineral (rancieite (Ca,Mn)Mn4O9) contained in a sample of Yucca Mountain tuff. However, Powell et al. (17) analyzed the same sample a few years later and showed that the Pu(V/VI) had reduced to Pu(IV). Using glass beads and pyrolusite, they proposed that the greater stability of the hydrolyzed-Pu(IV) surface species may in part be responsible for the presence of the reduced Pu species on oxidizing surfaces. Shaughnessy et al. (12) used XANES to show reduction of Pu(VI) by hausmannite and manganite. The objective of this study was to evaluate the impact of pH on aqueous Pu concentrations and oxidation state transformations using sorption and desorption studies. The desorption studies were conducted with a sediment collected from the Savannah River Site, located near Aiken, South Carolina that had been in contact with Pu (from PuIV(NO3)4) for 24 years. This sediment sample was taken from a longterm Pu transport study conducted in a field lysimeter (18). During the study, Pu had traveled 10 cm, with >95% of the Pu remaining within 1.25 cm of the source. By assuming in a reactive transport model that all the Pu existed as Pu(IV), Pu mobility was significantly underestimated by a factor of 3.5. Thus, Pu desorption, albeit quite little, was occurring in the lysimeter sediment system at a rate greater than was expected based on the assumption that the Pu remained as Pu(IV) and linear adsorption-desorption processes were in place. It was concluded that understanding the causes for the very small amount of Pu desorption were very important for predicting Pu mobility. Desorption is commonly thought to be the rate limiting process controlling contaminant transport because desorption processes tend to be appreciably slower than sorption processes.
Materials and Methods Aqueous Pu concentrations and Pu oxidation state distributions were determined in both desorption and sorption experiments. The tests were conducted with sediments collected from the field lysimeter study described above (Table 1; 18, 19). The sediment is from the SRS vadose zone and its surface properties are controlled largely by the presence of Fe-(oxy)hydroxide coatings. Total iron in the sediment is 1.5% Fe, of which 0.5% is free Fe (dithionite extractable Fe), 0.3% is amorphous Fe oxides (oxalate extractable Fe) (Table 1). The dominant crystalline minerals that can be detected by XRD are goethite and hematite. The low Mn concentrations suggest that Mn-mineralogy is likely not an important factor for redox chemistry. The desorption study was conducted with a sediment contaminated from a PuIV(NO3)4 source disk (described in more detail below), and the sorption study was conducted with the same sediment, but from a lysimeter that did not receive any Pu source material. The properties of the sediment for the desorption studies are reported in Table 1. The sediment properties for the sorption studies are the same except the Pu concentrations are at background levels, which are ∼0.10 Bq g-1 239/240Pu and HIV 15,888 5346 2925 630 3.9
a Data presented in Kaplan et al. (18) and measured by the wetchemistry method involving the parallel extraction of Pu into 0.5 M TTA in cyclohexane at pH 0.5 and 0.5 M HDEHP in heptane at pH 0.5, as described in Powell et al. (10) and in the Supporting Information. b Relative concentrations of minerals in clay-size ( 18 MΩ‚cm) to produce a range of pH levels. In parallel, 36 control tubes, containing solution but not sediment, were also set up to permit measuring proton activity in the absence of sediment. The pH of the suspensions and the control solutions were measured after a two week equilibration period and the PZSE calculated graphically, as outlined by Zelazny and Liming (20). Pu(V) Sorption Kinetics. The experimental design included nine contact times between 0 and 3000 min (50 h) and two replicates, for a total of 18 tubes. A 0.3 mL aliquot of a 0.5 µM 239Pu(V) working solution was added to each tube containing 10 mL of 0.02 M NaCl and 0.250 g of control (Pufree) sediment (24,300 ( 400 mg L-1). The working solution contained 98 ( 3% Pu(V) at time zero and 99 ( 6% Pu(V) at the conclusion of the experiment. The samples were mixed in the dark for up to 50 h. These experiments, as well as all others discussed here, were conducted on the lab benchtop and no effort was made to restrict the amount of CO2(g) permitted to come into contact with the various experimental systems. Control samples containing no sediment were prepared to monitor the stability of Pu(V) over the course of the experiment. No change in Pu oxidation state was detected in these controls. At the end of the study, the sediment was separated from the aqueous phase by centrifugation and passing through a 30 000 molecular-weight-cutoff membrane
filters (30k MWCO filters, ∼12-nm; Omega Microsep Centrifugal Device; Pall Corporation, East Hills, NY). No Pu adsorption to the filters was observed. The aqueous phase was then analyzed for total Pu concentration and Pu oxidation state distribution by the TTA/HDEHP (thenolytrifluoroacetone/bis(ethyhexyl)-phosphoric acid) solvent extraction method, as described below. The sediment at the end of the experiment was analyzed for Pu oxidation state analysis by the method described below. Pu(IV) and Pu(V) Working Solution Preparation. Experiments were performed with 239Pu(IV) and 239Pu(V)working solutions prepared from a 33.6 µM 239Pu(NO3)4 stock solution (Isotope Products, Valencia, CA). The solutions contained a small amount of 241Pu, which was easily separated from 239Pu during liquid scintillation counting and did not significantly add to the mass of Pu in the system. The 241Am generated through 241Pu decay accounted for less than 0.01% of the total alpha activity and was not considered significant. Pu(IV) working solutions were prepared by evaporating an aliquot of Pu(NO3)4 stock to incipient dryness several times in 1.0 M HNO3. The final residue was brought up in a small volume of 1.0 M HNO3 and diluted to 4.75 µM using deionized H2O. 239Pu(V) working solutions were prepared using the method described by Satio et al. (24). The Pu(IV) and Pu(V) working solutions were analyzed by the Pu oxidation distribution method and found to have the following compositions: the Pu(IV) spike was 98 ( 4% Pu(IV), 0 ( 0% Pu(V), and 3 ( 6% Pu(VI); the Pu(V) spike was 0 ( 0% Pu(IV), 98 ( 4% Pu(V) and 2 ( 1% Pu(VI). Pu(IV) and Pu(V) Sorption as a Function of pH. A 0.5 g aliquot of control (Pu-free) sediment was added to centrifuge tubes containing 10 mL of NaCl/HCl/NaOH solutions with similar ionic strength but varying pH levels (as described in detail in the Supporting Information). Two sets of sediment suspensions were prepared, one for Pu(IV) and the other for Pu(V). A 0.33 mL aliquot from a 15.7 µM Pu(V) working solution was added to one set of tubes to produce an initial Pu(V) concentration of 0.53 µM. Similarly, a 0.33 mL aliquot from an acidic 4.7 µM Pu(IV) working solution was added to the second set of tubes to achieve and initial Pu(IV) concentration of 0.15 µM. The Pu amended samples were placed on a platform shaker in the dark for 24 h. An aliquot of the aqueous phase was then filtered through a 30 000 molecular-weight-cutoff centrifugal filters (30 000 MWCO filters) and a 1.0 mL aliquot of the filtrate was removed for total Pu measurements. The Pu(V) set of tubes was returned to the platform shaker and permitted to equilibrate in the dark for an additional 32 days prior to measuring total Pu. Plutonium Desorption as a Function of pH. The objective of this study was to evaluate the desorption of Pu as a function of pH, with particular attention directed at the impact of pH on the aqueous phase Pu oxidation state. As mentioned above, the sediment sample was in contact with Pu (originally PuIV(NO3)4) for 24 years and as such represents a well-aged sample. The sample contained 62 880 Bq g-1 (1.56 × 10-4 mol kg-1) and the Pu existed on the sediment almost exclusively in the +4 oxidation state (Table 1). The red sandyclay-loam subsurface sediment had low organic matter concentrations, and its clay-size mineralogy was dominated by kaolinite. The experimental design included 12 pH values and two replicates for a total of 24 treatments. A 0.5 g sediment aliquot was placed in each of a series of test tubes containing 10 mL of NaCl/HCl/NaOH solutions of similar ionic strength but different H+ and OH- concentrations, as described in the Supporting Information. The suspensions were placed on a slow moving platform shaker for two weeks. The pH was then measured, and the solid and aqueous phases were separated by filtration through 30 000 MWCO filters. The aqueous phase was then further analyzed for total Pu
concentration and Pu oxidation state distribution by the TTA/ HDEHP solvent extraction method, as described below. Analytical Methods. Total Pu concentrations were measured with an alpha-beta discriminating liquid scintillation counter (Wallac Inc., model 1415, Boston, MA). Aliquots were mixed with HiSafe3 scintillation cocktail (Perkin-Elmer Life Sciences, Boston, MA) for liquid scintillation measurements. The liquid scintillation data was corrected for the background count rate obtained by measuring samples containing each matrix encountered during the experiments (i.e., TTAcyclohexance, HEDPA-heptane, HNO3, HCl, and NaCl). Additionally, alpha-beta discrimination was employed to lower the beta character of the background. In all cases, determination of Pu concentration and oxidation state was performed on the filtrate obtained after passing samples through 30 000 MWCO membranes. The reported error for all Pu measurements was propagated from liquid scintillation counting statistics. Preliminary studies with positive controls (spiked solution without sediment) showed extensive adsorption of Pu(IV), but not Pu(V), to the vial walls. Therefore, the vials holding the Pu(IV)-soil samples were immediately analyzed for the adsorption to the vial walls. Three 5-mL aliquots of deionized H2O were used to wash out the sediment and aqueous phase from each tube. After 5 mL of 1.0 M HNO3 was added to each tube, the tubes were weighed and then shaken for 1 h. A 1.0 mL aliquot of each acid solution was removed to determine the total Pu adsorbed to the vial walls. Analyses of these acid wash solutions indicated that between 0.2 and 1.0% of the added Pu(IV) adsorbed to the vial walls for samples containing sediment. Therefore, adsorption to vial walls was not a relevant factor for aqueous Pu(IV) removal during these experiments. Plutonium oxidation state distribution analysis was conducted on the sediment sample used in the kinetic study. The analysis was conducted on the duplicate sediment samples immediately after completion of the 50 h experiment. The procedure, validation of the procedure, and the quality control used to measure Pu oxidation state distribution in the sediment is described in detail by Powell et al. (10). This procedure employs ultrafiltration and solvent extraction techniques initially developed by Keeney-Kennicutt and Morse (4), and subsequently modified by Morgenstern and Choppin (15), Neu et al. (7), and Powell et al. (10). It quantifies the oxidation state distribution in both the solid and aqueous phase. A brief description of this technique follows (a full description is provided in the Supporting Information). Oxidation state analysis of aqueous phase Pu was performed by parallel extraction of the Pu solution into 0.5 M TTA in cyclohexane at pH 0.5 and 0.5 M HDEHP in heptane at pH 0.5 (7, 22, 23). A 0.5 M TTA solution in cyclohexane extracts Pu(IV) from a pH ≈ 0 aqueous phase, leaving Pu(V) and Pu(VI) behind. A 0.5 M HDEHP solution in heptane extracts Pu(IV) and Pu(VI) from a pH ≈ 0 aqueous phase, leaving Pu(V) behind. To quantify Pu oxidation state distributions in the sediment, the sediment was acidified to pH 1.5 using HClO4 for 15 min. The acid extract was filtered (30 000 MWCO filters) and analyzed for Pu oxidation state distribution by parallel solvent extractions into 0.5 M TTA in cyclohexane at pH 0.5 and 0.5 M HDEHP in heptane at pH 0.5 (4, 7, 10, 15). This fractionation scheme quantifies the Pu in the +4, +5, and +6 oxidation states (4, 7, 10, 15). This procedure can also detect Pu(III), but it was elected not to measure this oxidation state because previous work with this sediment has shown the presence of negligible concentrations,
